Algorithmic Impact Assessment as a Control System of Automated Disposition
Bibliographic record
Abstract
The Framework Act on Administration, scheduled to take effect in March 2021, provides an explicit basis for fully automated administrative decisions by introducing the concept of ‘automatic disposition’ in Article 20. Traditionally, the ‘automation of administration’ was significant as an auxiliary administrative procedure consisting of automatic mechanical devices such as traffic signal transmission, tax, and utility bill calculation. Moreover, since automated administrative decisions using artificial intelligence (AI) technology are also included in this automatic disposition, there is room for legal acceptance of automated administrative decisions based on artificial intelligence algorithms that exceed the level of conventional computerization or partial automation. However, it is still a long way from implementing a fully automated administrative decision system based on current law and having technical and institutional safety devices for it. On the other hand, the Framework Act on Intelligent Informatization, which took effect in December 2020, introduced a new ‘social impact assessment’ for intelligent information services that have far-reach effects on citizens’ lives. According to Article 56 of the Act, the main goal of this impact assessment system is to investigate and evaluate the impact of intelligent information services on society, economy, culture, etc., disclose the results, reflect them in policies, and directly recommend appropriate measures. Since the scope of intelligent information services subject to evaluation is not limited to the private sector, it should be considered that intelligent information services in the administrative sector can also be subject to social impact assessment in this article. It is reasonable to assume that automatic administrative services by artificial intelligence system belong to a wide range of intelligent information services prescribed by law, and that automatic disposition issued as a fully automated system based on intelligent information technology can also be subject to social impact assessment. Although no provisions have been found in law to assess the impact of automated administrative decisions, the government of Canada has implemented the “Directive on Automated Decision-Making” since April 2019, which systematically regulates the automated administrative decision-making system. The requirement that constitutes an important axis of this directive is the ‘algorithmic impact assessment'. For an automated administrative decision-making system that bears relatively stricter legality and public responsibility than private companies' services, it is necessary to process and operate an impact assessment that reviews safety, transparency, and solidity in advance. In this paper, referring to the Canadian case, as a control system of the algorithm-based administrative decision-making system, I propose the institutionalization of the “algorithmic impact assessment” in consideration of the connection with the social impact assessment under the current law.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.008 | 0.029 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.005 | 0.002 |
| Science and technology studies | 0.001 | 0.005 |
| Scholarly communication | 0.008 | 0.005 |
| Open science | 0.002 | 0.004 |
| Research integrity | 0.001 | 0.003 |
| Insufficient payload (model declined to judge) | 0.007 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".